Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chess Tiger versus Gulko

Author: James T. Walker

Date: 13:54:28 03/20/02

Go up one level in this thread


On March 20, 2002 at 16:44:24, pavel wrote:

>On March 20, 2002 at 16:32:05, James T. Walker wrote:
>
>>On March 20, 2002 at 16:27:02, Jonas Cohonas wrote:
>>
>>>On March 20, 2002 at 16:14:25, James T. Walker wrote:
>>>
>>>>On March 20, 2002 at 16:01:24, pavel wrote:
>>>>
>>>>>On March 20, 2002 at 15:35:13, James T. Walker wrote:
>>>>>
>>>>>>On March 20, 2002 at 14:54:44, pavel wrote:
>>>>>>
>>>>>>>On March 20, 2002 at 13:07:55, James T. Walker wrote:
>>>>>>>
>>>>>>>>On March 20, 2002 at 10:26:08, Jonas Cohonas wrote:
>>>>>>>>
>>>>>>>>>On March 20, 2002 at 09:43:45, James T. Walker wrote:
>>>>>>>>>
>>>>>>>>>>On March 20, 2002 at 06:24:14, Steve Maughan wrote:
>>>>>>>>>>
>>>>>>>>>>>The version of HIARCS used is the new one - HIARCS 8, which may well be up with
>>>>>>>>>>>the top engines.  It has been planned for release any time now.  According to
>>>>>>>>>>>sources close to HIARCS the ChessBase company rates the new HIARCS engine second
>>>>>>>>>>>only to Fritz 7.  We'll have to see but it has been a while since HIARCS 7.32
>>>>>>>>>>>was launched (May 99) and Mark Uniacke could have made quite a bit of progress
>>>>>>>>>>>in the last three years.
>>>>>>>>>>>
>>>>>>>>>>>Steve
>>>>>>>>>>
>>>>>>>>>>" ChessBase company rates the new HIARCS engine second only to Fritz 7."
>>>>>>>>>>
>>>>>>>>>>That's interesting.  I didn't know that.  I think it's funny since I rate Fritz
>>>>>>>>>>7 second only to Chess Tiger 14.0.  Never the less, I'm looking forward to the
>>>>>>>>>>new Hiarcs.
>>>>>>>>>>Jim
>>>>>>>>>
>>>>>>>>>Chesstiger 14.0 is very strong, but all my testings and others i have seen too,
>>>>>>>>>indicates that Fritz 7b is actually the strongest of the two.
>>>>>>>>>
>>>>>>>>>Regards
>>>>>>>>>Jonas
>>>>>>>>
>>>>>>>>Hello Jonas,
>>>>>>>>You may be right but,  I believe most of the testing you are referring to was
>>>>>>>>done on one computer with ponder off.  In my opinion the only thing you prove
>>>>>>>>with this setup is which is stronger in this setup.  This setup is not a "real
>>>>>>>>world" situation.  No programs actually compete this way except in these home
>>>>>>>>test.  I am basically referring to "Blitz" time controls since that is where I
>>>>>>>>have the most "experience".  In my database Chess Tiger 14.0 is still about 8
>>>>>>>>points ahead of Fritz 7 and both have played 1300/1400 + games vs various
>>>>>>>>opponents using two Athlon computers/auto232.  In my "Action" database which is
>>>>>>>>mostly game/25 or game/30 minutes Chess Tiger is leading Fritz 7 by 19 ELO but I
>>>>>>>>have less than 200 games each.  In my "Standard" database Fritz 7 is leading CT
>>>>>>>>14 by more than 50 points but again I only have150/200 games each.  These
>>>>>>>>databases where the number of games is low is very volatile since the next
>>>>>>>>series of games by either program tends to change the rating by fairly large
>>>>>>>>amounts.  Basically I'm leaving the "longer" time control testing up to SSDF
>>>>>>>>since they can do much more than I can.  The bottom line for me is that they are
>>>>>>>>so close that we may never know for sure which is strongest.  But it's OK for us
>>>>>>>>to have different opinions and its understandable since we have different
>>>>>>>>experiences.
>>>>>>>>Regards,
>>>>>>>>Jim
>>>>>>>
>>>>>>>I have seen many test set up in many types of machine including blitz and
>>>>>>>standard, in one or two machine and with ponder=on and off, in many rating
>>>>>>>lists. I perticipate in all top ICS, (iregularly), and from the results on fritz
>>>>>>>ICS server from several 100s opponents playing with several humdreds setups with
>>>>>>>fritz7, I can only say fritz7 is better than shredder6 as well as it is than
>>>>>>>chesstiger 14.
>>>>>>>
>>>>>>>this is a "fact".
>>>>>>>
>>>>>>>you might want to check up this thread.
>>>>>>>http://www.talkchess.com/forums/1/message.html?218204
>>>>>>>
>>>>>>>;)
>>>>>>>
>>>>>>>pavs
>>>>>>
>>>>>>I read it and see no proof that Fritz 7 is better than CT14.  It is certainly
>>>>>>much better than Shredder 6 (any version) at blitz.  I believe my 1400 games on
>>>>>>auto232 is more proof than you show.  It says there is only 7 points difference
>>>>>>between CT14 and F7 at blitz.  Show me your 1400 games and I will compare.  All
>>>>>>my games are on equal hardware (2 Athlon 1.4G machines) with ponder on.  Still
>>>>>>as I said in the other post everyone is entinled to their opinion.  I just
>>>>>>believe my opinion is based on more facts than others.
>>>>>>Jim
>>>>>
>>>>>
>>>>>I am not (not anymore, but used to) interested in  blitz games, so sorry to tell
>>>>>you that I wont go through the trouble to test it. I was not only referring to
>>>>>blitz but overall impression, even in blitz games perhaps your list (which I am
>>>>>yet to see), is the only list I have heard of which has chesstiger better than
>>>>>fritz7b.
>>>>>
>>>>>so do you have 1400 games between CT14 and Fritz7b (<-- this version I am
>>>>>refferring to, not all the jumbled up version of fritz7)?
>>>>>or 1400 games played seperatedly by CT14 and fritz7b
>>>>>or list of 1400games were ct14 and fritz7b exists.
>>>>>
>>>>>which one?
>>>>>
>>>>>pavs
>>>>
>>>>Hello Pavs,
>>>>My database (Blitz) consist of about 6000 games now.  Of that amout Fritz 7 has
>>>>played abou 1425 games vs various opponents.  CT14 has played about 1350 games
>>>>vs same opponents.  They have played against each other about 240 games which
>>>>has CT14 leading by about 20 ELO in those 240 games.  In the overall database
>>>>CT14 leads by only 8 ELO.  So as you can see that is my experience and it is
>>>>really too close to call but I certainly see no reason to suspect that F7 is
>>>>stronger than CT14.
>>>>Jim
>>>
>>>Program                          Elo    +   -   Games   Score   Av.Op.  Draws
>>>
>>>  1 Fritz 7                        : 2677   18  18   993    61.9 %   2593   34.1
>>>%
>>>  2 Chess Tiger 14.0               : 2654   19  16   978    57.9 %   2599   38.3
>>>%
>>>  3 Deep Fritz                     : 2641   20  17   999    56.2 %   2598   33.4
>>>%
>>>  4 Gambit Tiger 2.0               : 2631   20  15   996    55.2 %   2595   39.9
>>>%
>>>  5 Junior 7                       : 2623   21  17   983    54.2 %   2594   31.7
>>>%
>>>  6 Deep Junior 7                  : 2598   16  22   992    49.8 %   2599   32.0
>>>%
>>>  7 Shredder 6                     : 2598   16  22   992    49.2 %   2603   32.4
>>>%
>>>  8 Hiarcs 7.32                    : 2572   16  20   996    45.8 %   2601   34.2
>>>%
>>>  9 GandalfUCI                     : 2553   50  52   145    42.4 %   2606   25.5
>>>%
>>> 10 Nimzo 8                        : 2544   17  19   987    41.4 %   2604   33.9
>>>%
>>> 11 Crafty 18.13                   : 2542   20  21   779    40.9 %   2606   32.3
>>>%
>>> 12 Crafty 18.12                   : 2532   37  38   228    38.2 %   2616   35.1
>>>%
>>> 13 Goliath Light 1.5              : 2527   27  25   544    38.5 %   2609   26.7
>>>%
>>> 14 gandalf_432h                   : 2518   47  43   184    37.5 %   2607   27.2
>>>%
>>>
>>>(1) Fritz 7                   : 993 (+445,=339,-209), 61.9 %
>>>
>>>Chess Tiger 14.0              : 107 (+ 33,= 45,- 29), 51.9 %
>>>Crafty 18.12                  :  23 (+  9,= 12,-  2), 65.2 %
>>>Deep Fritz                    : 105 (+ 39,= 34,- 32), 53.3 %
>>>Shredder 6                    : 124 (+ 65,= 40,- 19), 68.5 %
>>>Deep Junior 7                 :  98 (+ 40,= 31,- 27), 56.6 %
>>>Nimzo 8                       :  85 (+ 40,= 31,- 14), 65.3 %
>>>Junior 7                      :  92 (+ 39,= 37,- 16), 62.5 %
>>>Gambit Tiger 2.0              :  83 (+ 33,= 33,- 17), 59.6 %
>>>Hiarcs 7.32                   :  80 (+ 42,= 24,- 14), 67.5 %
>>>Crafty 18.13                  :  92 (+ 48,= 30,- 14), 68.5 %
>>>Yace 0.99.56                  :  11 (+  4,=  2,-  5), 45.5 %
>>>Nejmet 3.0                    :   3 (+  3,=  0,-  0), 100.0 %
>>>Goliath Light 1.5             :  63 (+ 37,= 13,- 13), 69.0 %
>>>gandalf_432h                  :  15 (+  8,=  3,-  4), 63.3 %
>>>GandalfUCI                    :  12 (+  5,=  4,-  3), 58.3 %
>>>
>>>Regards
>>>Jonas
>>
>>Is it not true that the above database was done on one computer with ponder off?
>
>
>I doubt it was, even if it did, I dont think it will make much of a differance.
>
>pavs

Well in my opinion it does.  Also I don't think the above 107 games are more
important than my 240 games in head to head fighting.  And as the above also
shows there is very little difference.  So not much to disagree about except
which one is actually on top.  The data does not prove either one is best.  In
any case I expect the next Tiger will make it clear.
Jim



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.